New diagnostics and methods of assessing pregnant women at risk of cytomegalovirus

Human cytomegalovirus (CMV) infection can occur in pregnant women by primary infection or by non-primary infection, namely by either reactivation of the latent virus or reinfection with a different strain¹. In all cases the mother can transmit the virus to the fetus through the placenta^2,3. In the diagnosis of primary CMV infection, the gold standard is maternal seroconversion to CMV-specific antibodies. Currently, women are not routinely screened for CMV before conception or during pregnancy, thus CMV seroconversion is infrequently documented¹. Lastly, serological diagnosis of non-primary CMV infection is very difficult and very often unreliable since no optimal diagnostic methods are currently available. Today, the fetal compartment can be only studied by amniocentesis and ultrasound examination for the diagnosis and prognosis of CMV infection and generally, invasive diagnostic protocol can be only suggested to pregnant women with evidence of primary CMV infection acquired early in gestation and in case of abnormal findings suggestive of congenital infection¹. Therefore, a correct maternal diagnosis makes so that invasive prenatal diagnosis is only offered in selected cases. This report points out how a CMV-screening program combined with an advanced diagnostic protocol performed on pregnant women could identify those at high risk of transmitting the virus to their fetus. Furthermore, we evaluated the possible role of soluble HLA-G (sHLA-G) molecules detected in maternal and fetal samples in order to more accurately assess a greater risk of CMV-transmission and fetal/neonatal injury.

Testing for anti-CMV IgM antibodies is the most widely used and appropriate procedure for screening pregnant women⁴. Anti-CMV IgM antibodies are a good indicator of acute or recent infection, however it is not always correlated with active infection^4–6. The rate of CMV-IgM detection by screening test ranged from 3–5.7%^7–9, however only 7.5% of IgM-positive women have a congenitally infected fetus/newborn¹. Consequently, all pregnant women with a positive screening CMV-IgM test should be offered advanced diagnostic testing as early as possible in pregnancy (before week 12–16 of gestation). Anti-CMV IgG avidity testing^10–13, CMV-IgG and IgM immunoblotting (IB)^14–16, microneutralisation assay¹⁷, and detection of viral DNA in blood, saliva and urine samples^18–20 with Real Time PCR are currently the most reliable advanced procedures in order to identify all pregnant women who can transmit CMV infection.

Since 1994, pregnant women with test results showing seroconversion or IgM screening positivity were referred to our centre for further analysis¹⁹. In 2014, we performed a check up on 194 pregnant women during a 6 month period. Out of these 194 cases 104 (53.6%) did not know their CMV serostatus before pregnancy. At the time of recruitment, the patients were in the first or second trimester of pregnancy, except 28 who were in the third trimester (range 6–38 weeks gestation, median 14).

We tested blood, saliva and urine samples obtained from all 194 women and were able to identify five different groups using serological and virological advanced diagnosis. In the first group of 15 non-immune pregnant women, all samples were positive/borderline for CMV-IgG or IgM with the screening assays. After using the advanced tests, we obtained in all cases negative results with IB-CMV IgG and IgM antibodies, undetectable IgG-avidity and CMV-DNA negative in all body fluids.

In the second group of 68 patients with past infection, the screening tests identified borderline/positive results for IgM antibodies in all 68 cases. After using advanced testing, we obtained negative results in all cases for IB-IgM antibodies and CMV-DNA in all body fluids. Moreover, we found high avidity in all cases.

In the third group, advanced diagnosis was able to identify 57 pregnant women with primary CMV infection. Also in this group the screening test was IgM positive or borderline and the IB confirmed this specific IgM-positivity for CMV. IgG-avidity in all 57 samples was low/moderate and in 52 out of 57 patients (91.2%) viral DNA in whole blood, saliva and urine samples was detected.

In particular, we found CMV-DNA in 35 whole blood samples out of 52 (61.4%) and the number of copies ranged from <500 to 8700 copies/mL. Higher rates of positivity were detected in the saliva and urine samples, 75.4% and 64.9%, respectively. The range varied from <500 to a maximum of 44 000 copies/mL of saliva and <500 to 9700 copies/mL of urine. Positive viral detection and viral load in whole blood, saliva and urine were not associated with a greater risk of infection and/or fetal/neonatal injury.

The fourth group included 43 pregnant women with non-primary CMV infection. IB confirmed the positive results for IgM in 40 out of 43 cases; in the remaining three patients we were able to prove the diagnosis of non-primary CMV infection with the detection of viral DNA in urine and saliva. In 43 pregnant women with non-primary CMV infection, we observed a good sensitivity of virological tests in saliva and urine samples (48.8% and 41.9%, respectively) and very low sensitivity in whole blood samples (9.3%). When considering overall virological results, we found viral DNA in body fluids in 65% of patients with non-primary CMV infection. The DNA levels were very low in all body fluids, ranging from <500 to 900 copies/mL in urine and less than 500 copies/mL in both whole blood and saliva.

Finally in the last group, the advanced diagnosis investigation confirmed active CMV infection, however we were not able to identify the kind of infection, hence the reason why this group included 11 pregnant women with undefined CMV infections. The incidence of congenital CMV infection in the 5 groups of pregnant women classified with advanced diagnostic protocol is shown in Figure 1.

Figure 1. Prevalence of congenital CMV infection in 194 pregnant women at risk of transmitting the virus after serological screening.

Although generally the diagnosis of CMV infection remains complex, major goals have been achieved in recent years including maternal diagnosis with serological and virological tests. In particular, the use of advanced serological diagnosis has proven to be reliable in assessing pregnant women at risk of CMV infection. Likewise, virological diagnosis is also reliable and can support the serological diagnosis of primary, past and undefined CMV infection, as well as playing a role in the diagnosis of non-primary CMV infection.

In order to improve the identification of i) pregnant women who transmit the virus to their fetus and ii) CMV-infected and compromised fetuses, we studied the expression of soluble isoform of HLA-G (sHLA-G) during CMV infection in maternal blood and amniotic fluid samples.

HLA-G is a non-classical HLA class I antigen characterised by a low allelic polymorphism, compared with the HLA class^21,22. The HLA-G antigen is a tolerogenic molecule that acts on cells of both innate and adaptive immunity^23,24. Interestingly, HLA-G expression by cytotrophoblasts is down-modulated by CMV infection²⁵, while it is up-modulated in peripheral blood cells, with possible functional consequences in pregnancy immuno-regulation^26,27.

In this study, sHLA-G levels in serum and amniotic fluid samples were assayed in triplicate as previously reported^28,29, using an enzyme-linked immunosorbent assay (ELISA) and the monoclonal antibody MEM-G9 (Exbio), which recognises HLA-G molecules, in β2-microglobulin associated form. The intra-assay coefficient of variation (CV) was 1.4% and the inter-assay CV was 4.0%; the limit of sensitivity was 1.0 ng/mL.

We have an interim analysis of a clinical prospective trial which is enrolling 400 pregnant women suspected at routine CMV testing to have active CMV infection. Here, we report the results obtained from a first cohort of 166 pregnant women. At the moment of serological-virological advanced diagnosis, we evaluated sHLA-G levels in 171 serum samples of 55 pregnant women with primary CMV infection, 31 with non-primary, 69 with past infection, and 11 CMV-uninfected. The median levels of sHLA-G in pregnant women with primary were higher in comparison with non-primary CMV infection (45.16 ng/mL v. 10.58 ng/mL, P = 0.005; Student’s t-test). Furthermore, we observed lower median levels of sHLA-G serum between past CMV infected and uninfected women (14.68 ng/mL and 6.71 ng/mL, respectively). When we analysed the levels of sHLA-G in plasma samples from 55 primarily infected pregnant, considering transmitter and non-transmitter mothers, we did not find any statistical correlation (P = 0.72; Student’s t-test).

Finally, we analysed 25 amniotic fluid samples collected during amniocentesis (20–21 weeks gestation)¹⁹ from pregnant women with primary CMV infection arising before 14 weeks gestation. The comparison of the levels of sHLA-G between transmitter and non-transmitter mother was not statistically significant (P = 0.38; Student’s t-test).

The limited sample size does not permit firm conclusions, however our preliminary results suggest that sHLA-G detected in maternal plasma samples might be an additional biomarker of CMV infection that could be considered in combination with currently used serological and virological markers.

The laboratory diagnosis of CMV infection proves to be a reliable tool, provided that pregnant women are checked from the first weeks of gestation. Moreover, the use of advanced serological and virological maternal tests allow clinicians to identify women who are at higher risk of transmitting CMV to their fetus; however, they do not identify the infected fetuses, therefore making it necessary to offer prenatal diagnosis.

Nevertheless, major limitations of prenatal diagnosis of CMV should be acknowledged; amniocentesis is an invasive procedure and positive results of amniotic fluid tests do not discriminate between infected fetuses and compromised fetuses. For this reason, researchers continue to work on the prognosis factors for the CMV disease. All in all, our very preliminary results in this study suggest that sHLA-G could be a sensitive marker in order to monitor CMV infection during pregnancy.